Stable Video Diffusion can generate AI videos on a 7 year old GPU

AI video generation has become increasingly popular in recent years, providing a powerful tool for creating realistic videos with minimal effort. Until recently, however, the computing power needed to generate high-quality AI videos was extremely expensive and mostly limited to cutting-edge research labs.

A new method called stable video diffusion (SVD) may have changed this narrative. A team of researchers from the University of California, Davis, developed SVD, which uses a 7-year-old NVIDIA GPU to generate AI videos without any additional hardware.

The team found that SVD outperforms existing methods such as Deep Video Portraits and Generative Adversarial Networks. In addition, SVD is more reliable, with videos generated on different machines looking almost identical. This means that the same video can easily be produced on different types of machines, regardless of their age.

Furthermore, the researchers demonstrated how SVD could be used to create “synthetic humans”, which are AI-generated characters that look and behave like real people. This opens up a variety of possibilities for both gaming and movies, as these synthetic humans could serve as controllable avatars or even star in films.

In summary, the development of SVD has made AI video generation much more accessible. With this new method, anyone with a 7-year-old NVIDIA GPU can produce high-quality AI videos effortlessly. The ability to generate synthetic humans adds an exciting new dimension to AI video generation, making it a powerful tool for filmmakers and game developers alike.

Read more here: External Link